Hey, unfortunately, I am no guru at the math behind the L.O. I would actually love to learn the math behind LO versus source state. I only know what the spec sheet I have says for the device and that it works as I have been using it for close to two years now... beyond that I can't tell you much. The numbers DO seem low to me, I expect it to at least be above 10 ghz.
But, I cant argue with a company that made me a working product to stack legacy into DP. I gave a company what I wanted to be designed as best I could explaining how I wanted the end result (the end band stack, putting both even and odd into 950-1450 chunks then band stacking one on top of the other to get an end 950 -2150 mhz) and they gave me those numbers, and a spec sheet and two prototypes that worked. I would be happy to take a picture as proof and share the sheet's specs with you...?
Maybe you can make better sense of those numbers than I can? ;0 Not trying to argue with you, I respect your skills more so than my own without question, but I also cant argue with what is written in front of me from an engineer that designed a working product? ;0
I would be happy to remove any trace of my name and PM you the design specs. drop me a PM sometime. I would be interested to actually figure the math out behind this. I believe the inversion I referenced is taking a chunk of 950-1450 (the IF in this instance) and putting it in the 1650 - 2150 range. The end result obviously can't be inverted. (It would end up in the upside down state you explained) The inversion instead references the original state versus the end state here. It needs manipulation to be put into the stack. The end state is an inversion of the source state.
-B